Login
- Title:
Dataset-level societal bias mitigation with text-to-image model
- Author(s):
- Yusuke Hirota, Jerone Andrews, Dora Zhao, Orestis Papakyriakopoulos, Apostolos Modas, Alice Xiang
- Abstract:
- Systems and methods are used to mitigate societal bias in image-text datasets by removing spurious correlations between protected groups and image attributes. Using text-guided inpainting models, the methods ensures protected group independence from all attributes and mitigates inpainting biases through data filtering. Evaluations on multi-label image classification and image captioning tasks show that the methods effectively reduce bias without compromising performance across various models.
- Published as:
- Patent # US20250148657A1
- Month:
- May
- Year:
- 2025
- Language:
- de
BibTeX